The present invention relates to a storage system, and more particularly to a lock control for a storage shared by a plurality of storage utilizing units, such as central processors, I/O processors, etc.
In a system in which a plurality of storage utilizing units are sharing a main storage and are operating independently, such as a multiprocessing system, during the time when one storage utilizing unit is accessing a certain storage area of the main storage, or during the time when there is a possibility of instruction retry being performed with respect to data within a certain storage area, the access to the storage area by the remaining storage utilizing units must be inhibited. Such access inhibition as described above is called "lock".
One of the storage lock controls according to the prior art is disclosed in the European Patent Publication No. 006l570 in which a lock bit is associated with each entry of a cache directory which belongs to each processor (CP) and the lock bit in the ON state or "1" state indicates that the main storage area represented by an address within the associated entry is locked. In this control system, when an access request is issued by one CP, the cache directories of the remaining CPs are searched in order to determine whether or not the requested address has been locked by other CPs. Accordingly, the cache directory of each CP additionally undergoes search operations caused by access requests from the other CPs. An access request from any one of the CPs can be accepted after the cache directories of the remaining CPs have been checked and after having confirmed that the requested address has not been locked by any other CPs.
If a special register for holding the locked address is prepared in each CP, separately from the cache directory, every cache directory is released from the search operation which is due to access requests from other CPs. In this case, however, all the special registers in the remaining CPs must be referred to before determining if the access request from a certain CP is allowed, thus leaving the turn around time unimproved. If the lock control is carried out by a single storage controller intervening between the shared storage and the plurality of CPs, with the whole locked addresses of the entire CPs being held in the controller, the problem mentioned above may be overcome. In a system including a plurality of storage controllers, however, similar problems as described above will occur with respect to these storage controllers.